README.md

In richysix/geneExpr: A Shiny App For Visualising RNA-seq Data

geneExpr

A shiny app for visualising gene expression data. https://richysix.shinyapps.io/geneexpr/

Instructions

The app is designed to display gene expression data produced by RNA-seq.

Loading data

Data can be loaded using the buttons at the top of the side bar control panel. Users need to supply a sample file along with a count file.

Count File

The count file should be a tab-separated file containing genes in rows and samples in columns. The 'Gene ID' and 'Gene Name' columns are required. If you are using IDs that don't obviously correspond to a gene name, simply duplicate the IDs column. The other required columns are the counts for each different sample either as already normalised counts or raw counts. If normalised count columns are provided these will be used directly (The column names must match the sample IDs with ' normalised count' appended). Otherwise count columns will be loaded into DESeq2 and normalised for library size (In this case the column names must match the sample IDs with ' count' appended). The normalised counts are then displayed. If both normalised and unnormalised counts are provided, the normalised counts will be used.

e.g.

Gene ID Gene Name 012525 count 012526 count 012536 count 012537 count 012525 normalised count 012526 normalised count 012536 normalised count 012537 normalised count ENSDARG00000000212 krt97 10 12 35 42 12.3 14.6 35.6 45.1 ENSDARG00000000567 znf281a 345 333 357 365 322.5 343.7 363.2 380.0

Sample File

The sample file is a tab-separated file containg at least two columns. The required columns are the sample IDs that match the (normalised) count column names in the count file. This must be the first column. The other required column should be labelled "condition" and details how the samples are divided into groups (e.g. Control and Treated). If the file contains a column labelled "sampleName" these names will be displayed on the heatmap instead of the IDs from the first column. The count columns will be reordered based on the sample file.

e.g.

condition sampleName batch 012536 Control Ctrl1 A 012537 Control Ctrl2 B 012525 Treated Trt1 A 012526 Treated Trt2 B

Subset by Gene id

A text file of Ensembl gene ids to subset the heatmap to can be uploaded using this button. Any ids that can't be matched will be listed in a warning alert above the heatmap.

Heatmap

The default plot is a heatmap showing all the genes and all the samples in the count data. Users can zoom in to a section of the plot by selecting an area of the plot and double-clicking on it. Double-clicking without highlighting resets the view back to all genes and all samples.

Filter Genes

The genes can be filtered according to the mean counts across all samples. Both minimum and maximum thresholds are available to set with the sliders.

Transform counts

The counts can be transformed as follows

• Raw - The untransformed normalised counts (Default)
• Max Scaled - Each row is scaled to the maximum value for that row.
• log10
• Mean Centred and Scaled - Each row is centred by subtracting the mean and scaled by dividing by the standard deviation for that row.

Clustering

Both the rows (genes) and columns (samples) of the heatmap can be clustered. Currently, the clustering is done by hierarchical clustering of the Pearson correlation coefficients between genes/samples. Only the genes/samples currently displayed in the heatmap are used for the clustering

Axis labels

Labels are added to the x axis if there are less than 48 columns and to the y axis if there are less than 80 rows. These can be removed using the Axis label checkboxes. X-axis labels will be the sample IDs from the first column of the samples file or from the "sampleNames" column if it exists. Y-axis labels will be taken from the "Gene Name" column of the counts file.

Fill Limits

The limits for the colour scale can be set manually. This can be useful for produces multiple heatmaps with the same colour scales.

Downloads

The Download plot button downloads the current plot as either a pdf, png, eps or svg. The wdith and height of the plot can be specified.

The Download Counts (tsv) button downloads the normalised count data for just the genes and samples displayed in the heatmap.

The Download Genes (txt) button downloads a file containing just the gene ids for the genes and samples displayed in the heatmap.

Load RData file

As an alternative to sample/count files, a pre-computed R data file containing a DESeq2 DESeqDataSet object (named DESeqData) can be uploaded using the Load .RData File button

Counts tab

The counts tab shows the normalised counts for the currently selected Genes and Samples. The Download Counts (tsv) button can be used to download the data in the table.

Transformed Counts tab

The Transformed Counts tab shows the transformed counts for the currently selected Genes and Samples. If no transformation has been applied, this will be the same as the Counts data. The Download Transformed counts (tsv) button can be used to download the data in this table.

Prerequisites

• shiny
• shinyjs
• shinyBS
• ggplot2
• DESeq2
• DT
• rnaseqVis
• rprojroot
• devtools

richysix/geneExpr documentation built on Oct. 7, 2022, 2:18 a.m.